spacehub

surshar / claude-code

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

Built by Anthropic · github.com/anthropics

claude-code / CHANGELOG.md
GitHub Actionschore: Update CHANGELOG.md 4392352 · on Dec 16, 2025 History
1,138 lines · 46.6 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.71
4
5- Added /config toggle to enable/disable prompt suggestions
6- Added `/settings` as an alias for the `/config` command
7- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
8- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
9- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
10- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
11- New syntax highlighting engine for native build
12
13## 2.0.70
14
15- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
16- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
17- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
18- Added `plan_mode_required` spawn parameter for teammates to require plan approval before implementing changes
19- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
20- Fixed input being cleared when processing queued commands while the user was typing
21- Fixed prompt suggestions replacing typed input when pressing Tab
22- Fixed diff view not updating when terminal is resized
23- Improved memory usage by 3x for large conversations
24- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
25- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
26- Fix thinking mode toggle in /config not persisting correctly
27- Improve UI for file creation permission dialog
28
29## 2.0.69
30
31- Minor bugfixes
32
33## 2.0.68
34
35- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
36- Fixed a bug where disallowed MCP tools were visible to the model
37- Fixed an issue where steering messages could be lost while a subagent is working
38- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
39- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
40- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
41
42## 2.0.67
43
44- Thinking mode is now enabled by default for Opus 4.5
45- Thinking mode configuration has moved to /config
46- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
47- Show reason why autoupdater is disabled in `/doctor`
48- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
49- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
50- Fixed scroll position resetting after deleting a permission rule in `/permissions`
51- Fixed word deletion (opt+delete) and word navigation (opt+arrow) not working correctly with non-Latin text such as Cyrillic, Greek, Arabic, Hebrew, Thai, and Chinese
52- Fixed `claude install --force` not bypassing stale lock files
53- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
54- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
55
56## 2.0.65
57
58- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
59- Added context window information to status line input
60- Added `fileSuggestion` setting for custom `@` file search commands
61- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
62- Fixed prompt not being saved to history when aborting a query with Escape
63- Fixed Read tool image handling to identify format from bytes instead of file extension
64
65## 2.0.64
66
67- Made auto-compacting instant
68- Agents and bash commands can run asynchronously and send messages to wake up the main agent
69- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
70- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
71- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
72- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
73- Fixed auto-loading .env when using native installer
74- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
75- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
76- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
77- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
78- Bedrock: Improve efficiency of token counting
79- Bedrock: Add support for `aws login` AWS Management Console credentials
80- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
81
82## 2.0.62
83
84- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
85- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
86- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
87- Fixed slash command selection not working when multiple commands share the same name
88- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
89- Fixed running versions getting removed because lock file incorrectly going stale
90- Fixed IDE diff tab not closing when rejecting file changes
91
92## 2.0.61
93
94- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
95
96## 2.0.60
97
98- Added background agent support. Agents run in the background while you work
99- Added --disable-slash-commands CLI flag to disable all slash commands
100- Added model name to "Co-Authored-By" commit messages
101- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
102- Updated Fetch to skip summarization for pre-approved websites
103- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
104
105## 2.0.59
106
107- Added --agent CLI flag to override the agent setting for the current session
108- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
109- VS Code: Fixed .claude.json config file being read from incorrect location
110
111## 2.0.58
112
113- Pro users now have access to Opus 4.5 as part of their subscription!
114- Fixed timer duration showing "11m 60s" instead of "12m 0s"
115- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
116
117## 2.0.57
118
119- Added feedback input when rejecting plans, allowing users to tell Claude what to change
120- VSCode: Added streaming message support for real-time response display
121
122## 2.0.56
123
124- Added setting to enable/disable terminal progress bar (OSC 9;4)
125- VSCode Extension: Added support for VS Code's secondary sidebar (VS Code 1.97+), allowing Claude Code to be displayed in the right sidebar while keeping the file explorer on the left. Requires setting sidebar as Preferred Location in the config.
126
127## 2.0.55
128
129- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
130- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
131- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
132- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
133
134## 2.0.54
135
136- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
137- Fix issue with excessive iTerm notifications
138
139## 2.0.52
140
141- Fixed duplicate message display when starting Claude with a command line argument
142- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
143- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
144- Permit some uses of `$!` in bash commands
145
146## 2.0.51
147
148- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
149- Introducing Claude Code for Desktop: https://claude.com/download
150- To give you room to try out our new model, we've updated usage limits for Claude Code users. See the Claude Opus 4.5 blog for full details
151- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
152- Plan Mode now builds more precise plans and executes more thoroughly
153- Usage limit notifications now easier to understand
154- Switched `/usage` back to "% used"
155- Fixed handling of thinking errors
156- Fixed performance regression
157
158## 2.0.50
159
160- Fixed bug preventing calling MCP tools that have nested references in their input schemas
161- Silenced a noisy but harmless error during upgrades
162- Improved ultrathink text display
163- Improved clarity of 5-hour session limit warning message
164
165## 2.0.49
166
167- Added readline-style ctrl-y for pasting deleted text
168- Improved clarity of usage limit warning message
169- Fixed handling of subagent permissions
170
171## 2.0.47
172
173- Improved error messages and validation for `claude --teleport`
174- Improved error handling in `/usage`
175- Fixed race condition with history entry not getting logged at exit
176- Fixed Vertex AI configuration not being applied from `settings.json`
177
178## 2.0.46
179
180- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
181
182## 2.0.45
183
184- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
185- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
186- Send background tasks to Claude Code on the web by starting a message with `&`
187
188## 2.0.43
189
190- Added `permissionMode` field for custom agents
191- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
192- Added skills frontmatter field to declare skills to auto-load for subagents
193- Added the `SubagentStart` hook event
194- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
195- Fixed duplicate rendering of some messages in the UI
196- Fixed some visual flickers
197- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
198
199## 2.0.42
200
201- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
202
203## 2.0.41
204
205- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
206- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
207- Fixed incorrect labeling of user settings vs project settings in command descriptions
208- Fixed crash when plugin command hooks timeout during execution
209- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
210- Fixed broken security documentation links in trust dialogs and onboarding
211- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
212- ctrl-r history search landing on a slash command no longer cancels the search
213- SDK: Support custom timeouts for hooks
214- Allow more safe git commands to run without approval
215- Plugins: Added support for sharing and installing output styles
216- Teleporting a session from web will automatically set the upstream branch
217
218## 2.0.37
219
220- Fixed how idleness is computed for notifications
221- Hooks: Added matcher values for Notification hook events
222- Output Styles: Added `keep-coding-instructions` option to frontmatter
223
224## 2.0.36
225
226- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
227- Fixed queued messages being incorrectly executed as bash commands
228- Fixed input being lost when typing while a queued message is processed
229
230## 2.0.35
231
232- Improve fuzzy search results when searching commands
233- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
234- Added `CLAUDE_CODE_EXIT_AFTER_STOP_DELAY` environment variable to automatically exit SDK mode after a specified idle duration, useful for automated workflows and scripts
235- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
236- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
237
238## 2.0.34
239
240- VSCode Extension: Added setting to configure the initial permission mode for new conversations
241- Improved file path suggestion performance with native Rust-based fuzzy finder
242- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
243- Fixed memory crash when reading or writing large files (especially base64-encoded images)
244
245## 2.0.33
246
247- Native binary installs now launch quicker.
248- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
249- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
250
251## 2.0.32
252
253- Un-deprecate output styles based on community feedback
254- Added `companyAnnouncements` setting for displaying announcements on startup
255- Fixed hook progress messages not updating correctly during PostToolUse hook execution
256
257## 2.0.31
258
259- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
260- Vertex: add support for Web Search on supported models
261- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
262- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
263- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
264- Fixed plugin uninstall not removing plugins
265
266## 2.0.30
267
268- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
269- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
270- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
271- Added prompt-based stop hooks
272- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
273- Enabled SSE MCP servers on native build
274- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
275- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
276- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
277- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
278- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
279- Fixed bug that saved session permissions to local settings
280- Fixed MCP tools not being available to sub-agents
281- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
282- Fixed delay when navigating through typeahead suggestions with arrow keys
283- VSCode: Restored selection indicator in input footer showing current file or code selection status
284
285## 2.0.28
286
287- Plan mode: introduced new Plan subagent
288- Subagents: claude can now choose to resume subagents
289- Subagents: claude can dynamically choose the model used by its subagents
290- SDK: added --max-budget-usd flag
291- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
292- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
293- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
294- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
295- Various other bug fixes
296
297## 2.0.27
298
299- New UI for permission prompts
300- Added current branch filtering and search to session resume screen for easier navigation
301- Fixed directory @-mention causing "No assistant message found" error
302- VSCode Extension: Add config setting to include .gitignored files in file searches
303- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
304
305## 2.0.25
306
307- Removed legacy SDK entrypoint. Please migrate to @anthropic-ai/claude-agent-sdk for future SDK updates: https://platform.claude.com/docs/en/agent-sdk/migration-guide
308
309## 2.0.24
310
311- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
312- Claude Code Web: Support for Web -> CLI teleport
313- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
314- Bedrock: Display awsAuthRefresh output when auth is required
315
316## 2.0.22
317
318- Fixed content layout shift when scrolling through slash commands
319- IDE: Add toggle to enable/disable thinking.
320- Fix bug causing duplicate permission prompts with parallel tool calls
321- Add support for enterprise managed MCP allowlist and denylist
322
323## 2.0.21
324
325- Support MCP `structuredContent` field in tool responses
326- Added an interactive question tool
327- Claude will now ask you questions more often in plan mode
328- Added Haiku 4.5 as a model option for Pro users
329- Fixed an issue where queued commands don't have access to previous messages' output
330
331## 2.0.20
332
333- Added support for Claude Skills
334
335## 2.0.19
336
337- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
338- Fixed a bug where Haiku was unnecessarily called in print mode
339
340## 2.0.17
341
342- Added Haiku 4.5 to model selector!
343- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
344- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
345- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
346- OTEL: support HTTP_PROXY and HTTPS_PROXY
347- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
348
349## 2.0.15
350
351- Fixed bug with resuming where previously created files needed to be read again before writing
352- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
353
354## 2.0.14
355
356- Fix @-mentioning MCP servers to toggle them on/off
357- Improve permission checks for bash with inline env vars
358- Fix ultrathink + thinking toggle
359- Reduce unnecessary logins
360- Document --system-prompt
361- Several improvements to rendering
362- Plugins UI polish
363
364## 2.0.13
365
366- Fixed `/plugin` not working on native build
367
368## 2.0.12
369
370- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
371- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
372- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
373- `/plugin validate` command for validating plugin structure and configuration
374- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
375- Plugin documentation available at https://code.claude.com/docs/en/plugins
376- Comprehensive error messages and diagnostics via `/doctor` command
377- Avoid flickering in `/model` selector
378- Improvements to `/help`
379- Avoid mentioning hooks in `/resume` summaries
380- Changes to the "verbose" setting in `/config` now persist across sessions
381
382## 2.0.11
383
384- Reduced system prompt size by 1.4k tokens
385- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
386- Fixed Opus fallback rate limit errors appearing incorrectly
387- Fixed /add-dir command selecting wrong default tab
388
389## 2.0.10
390
391- Rewrote terminal renderer for buttery smooth UI
392- Enable/disable MCP servers by @mentioning, or in /mcp
393- Added tab completion for shell commands in bash mode
394- PreToolUse hooks can now modify tool inputs
395- Press Ctrl-G to edit your prompt in your system's configured text editor
396- Fixes for bash permission checks with environment variables in the command
397
398## 2.0.9
399
400- Fix regression where bash backgrounding stopped working
401
402## 2.0.8
403
404- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
405- IDE: Add drag-and-drop support for files and folders in chat
406- /context: Fix counting for thinking blocks
407- Improve message rendering for users with light themes on dark terminals
408- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
409
410## 2.0.5
411
412- IDE: Fix IME unintended message submission with Enter and Tab
413- IDE: Add "Open in Terminal" link in login screen
414- Fix unhandled OAuth expiration 401 API errors
415- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
416
417## 2.0.1
418
419- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
420- Various bug fixes and presentation improvements
421
422## 2.0.0
423
424- New native VS Code extension
425- Fresh coat of paint throughout the whole app
426- /rewind a conversation to undo code changes
427- /usage command to see plan limits
428- Tab to toggle thinking (sticky across sessions)
429- Ctrl-R to search history
430- Unshipped claude config command
431- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
432- SDK: The Claude Code SDK is now the Claude Agent SDK
433- Add subagents dynamically with `--agents` flag
434
435## 1.0.126
436
437- Enable /context command for Bedrock and Vertex
438- Add mTLS support for HTTP-based OpenTelemetry exporters
439
440## 1.0.124
441
442- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
443- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
444- No longer inform Claude of the list of allowed tools when permission is denied
445- Fixed security vulnerability in Bash tool permission checks
446- Improved VSCode extension performance for large files
447
448## 1.0.123
449
450- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
451- Fixed thinking mode triggering on negation phrases like "don't think"
452- Fixed rendering performance degradation during token streaming
453- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
454- Enhanced BashTool environment snapshot logging
455- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
456- Migrated --debug logging to a file, to enable easy tailing & filtering
457
458## 1.0.120
459
460- Fix input lag during typing, especially noticeable with large prompts
461- Improved VSCode extension command registry and sessions dialog user experience
462- Enhanced sessions dialog responsiveness and visual feedback
463- Fixed IDE compatibility issue by removing worktree support check
464- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
465
466## 1.0.119
467
468- Fix Windows issue where process visually freezes on entering interactive mode
469- Support dynamic headers for MCP servers via headersHelper configuration
470- Fix thinking mode not working in headless sessions
471- Fix slash commands now properly update allowed tools instead of replacing them
472
473## 1.0.117
474
475- Add Ctrl-R history search to recall previous commands like bash/zsh
476- Fix input lag while typing, especially on Windows
477- Add sed command to auto-allowed commands in acceptEdits mode
478- Fix Windows PATH comparison to be case-insensitive for drive letters
479- Add permissions management hint to /add-dir output
480
481## 1.0.115
482
483- Improve thinking mode display with enhanced visual effects
484- Type /t to temporarily disable thinking mode in your prompt
485- Improve path validation for glob and grep tools
486- Show condensed output for post-tool hooks to reduce visual clutter
487- Fix visual feedback when loading state completes
488- Improve UI consistency for permission request dialogs
489
490## 1.0.113
491
492- Deprecated piped input in interactive mode
493- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
494
495## 1.0.112
496
497- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
498- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
499- Hooks: Added systemMessage support for SessionEnd hooks
500- Added `spinnerTipsEnabled` setting to disable spinner tips
501- IDE: Various improvements and bug fixes
502
503## 1.0.111
504
505- /model now validates provided model names
506- Fixed Bash tool crashes caused by malformed shell syntax parsing
507
508## 1.0.110
509
510- /terminal-setup command now supports WezTerm
511- MCP: OAuth tokens now proactively refresh before expiration
512- Fixed reliability issues with background Bash processes
513
514## 1.0.109
515
516- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
517
518## 1.0.106
519
520- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
521
522## 1.0.97
523
524- Settings: /doctor now validates permission rule syntax and suggests corrections
525
526## 1.0.94
527
528- Vertex: add support for global endpoints for supported models
529- /memory command now allows direct editing of all imported memory files
530- SDK: Add custom tools as callbacks
531- Added /todos command to list current todo items
532
533## 1.0.93
534
535- Windows: Add alt + v shortcut for pasting images from clipboard
536- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
537
538## 1.0.90
539
540- Settings file changes take effect immediately - no restart required
541
542## 1.0.88
543
544- Fixed issue causing "OAuth authentication is currently not supported"
545- Status line input now includes `exceeds_200k_tokens`
546- Fixed incorrect usage tracking in /cost.
547- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
548- Bedrock: Updated default Sonnet model to Sonnet 4
549
550## 1.0.86
551
552- Added /context to help users self-serve debug context issues
553- SDK: Added UUID support for all SDK messages
554- SDK: Added `--replay-user-messages` to replay user messages back to stdout
555
556## 1.0.85
557
558- Status line input now includes session cost info
559- Hooks: Introduced SessionEnd hook
560
561## 1.0.84
562
563- Fix tool_use/tool_result id mismatch error when network is unstable
564- Fix Claude sometimes ignoring real-time steering when wrapping up a task
565- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
566- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
567
568## 1.0.83
569
570- @-mention: Support files with spaces in path
571- New shimmering spinner
572
573## 1.0.82
574
575- SDK: Add request cancellation support
576- SDK: New additionalDirectories option to search custom paths, improved slash command processing
577- Settings: Validation prevents invalid fields in .claude/settings.json files
578- MCP: Improve tool name consistency
579- Bash: Fix crash when Claude tries to automatically read large files
580
581## 1.0.81
582
583- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
584- Agents: Fix custom agent loading when agent files are unparsable
585
586## 1.0.80
587
588- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
589
590## 1.0.77
591
592- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
593- SDK: Add session support and permission denial tracking
594- Fix token limit errors in conversation summarization
595- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
596
597## 1.0.73
598
599- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
600- MCP: Press Esc to cancel OAuth authentication flows
601- Bash: Improved command validation and reduced false security warnings
602- UI: Enhanced spinner animations and status line visual hierarchy
603- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
604
605## 1.0.72
606
607- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
608
609## 1.0.71
610
611- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
612- Customizable status line: add your terminal prompt to Claude Code with /statusline
613
614## 1.0.70
615
616- Performance: Optimized message rendering for better performance with large contexts
617- Windows: Fixed native file search, ripgrep, and subagent functionality
618- Added support for @-mentions in slash command arguments
619
620## 1.0.69
621
622- Upgraded Opus to version 4.1
623
624## 1.0.68
625
626- Fix incorrect model names being used for certain commands like `/pr-comments`
627- Windows: improve permissions checks for allow / deny tools and project trust. This may create a new project entry in `.claude.json` - manually merge the history field if desired.
628- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
629- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
630- SDK: Added canUseTool callback support for tool confirmation
631- Added `disableAllHooks` setting
632- Improved file suggestions performance in large repos
633
634## 1.0.65
635
636- IDE: Fixed connection stability issues and error handling for diagnostics
637- Windows: Fixed shell environment setup for users without .bashrc files
638
639## 1.0.64
640
641- Agents: Added model customization support - you can now specify which model an agent should use
642- Agents: Fixed unintended access to the recursive agent tool
643- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
644- SDK: Fixed user input tracking across multi-turn conversations
645- Added hidden files to file search and @-mention suggestions
646
647## 1.0.63
648
649- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
650
651## 1.0.62
652
653- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
654- Hooks: Added SessionStart hook for new session initialization
655- /add-dir command now supports typeahead for directory paths
656- Improved network connectivity check reliability
657
658## 1.0.61
659
660- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
661- Settings: Added `--settings` flag to load settings from a JSON file
662- Settings: Fixed resolution of settings files paths that are symlinks
663- OTEL: Fixed reporting of wrong organization after authentication changes
664- Slash commands: Fixed permissions checking for allowed-tools with Bash
665- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
666- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
667- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
668
669## 1.0.60
670
671- You can now create custom subagents for specialized tasks! Run /agents to get started
672
673## 1.0.59
674
675- SDK: Added tool confirmation support with canUseTool callback
676- SDK: Allow specifying env for spawned process
677- Hooks: Exposed PermissionDecision to hooks (including "ask")
678- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
679- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
680
681## 1.0.58
682
683- Added support for reading PDFs
684- MCP: Improved server health status display in 'claude mcp list'
685- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
686
687## 1.0.57
688
689- Added support for specifying a model in slash commands
690- Improved permission messages to help Claude understand allowed tools
691- Fix: Remove trailing newlines from bash output in terminal wrapping
692
693## 1.0.56
694
695- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
696- Fixes for WSL IDE detection
697- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
698
699## 1.0.55
700
701- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
702- Windows: fixed Ctrl+Z crash
703- SDK: Added ability to capture error logging
704- Add --system-prompt-file option to override system prompt in print mode
705
706## 1.0.54
707
708- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
709- Custom slash commands: Added argument-hint to frontmatter
710- Windows: OAuth uses port 45454 and properly constructs browser URL
711- Windows: mode switching now uses alt + m, and plan mode renders properly
712- Shell: Switch to in-memory shell snapshot to fix file-related errors
713
714## 1.0.53
715
716- Updated @-mention file truncation from 100 lines to 2000 lines
717- Add helper script settings for AWS token refresh: awsAuthRefresh (for foreground operations like aws sso login) and awsCredentialExport (for background operation with STS-like response).
718
719## 1.0.52
720
721- Added support for MCP server instructions
722
723## 1.0.51
724
725- Added support for native Windows (requires Git for Windows)
726- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
727- Settings: /doctor can now help you identify and fix invalid setting files
728- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
729- Increased auto-compact warning threshold from 60% to 80%
730- Fixed an issue with handling user directories with spaces for shell snapshots
731- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
732- Custom slash commands: Fixed user-level commands in subdirectories
733- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
734
735## 1.0.48
736
737- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
738- Added progress messages to Bash tool based on the last 5 lines of command output
739- Added expanding variables support for MCP server configuration
740- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
741- Improved IDE extension path handling when Claude Code runs in WSL
742- Hooks: Added a PreCompact hook
743- Vim mode: Added c, f/F, t/T
744
745## 1.0.45
746
747- Redesigned Search (Grep) tool with new tool input parameters and features
748- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
749- Fixed config file corruption issue by enforcing atomic writes
750- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
751- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
752- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
753
754## 1.0.44
755
756- New /export command lets you quickly export a conversation for sharing
757- MCP: resource_link tool results are now supported
758- MCP: tool annotations and tool titles now display in /mcp view
759- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
760
761## 1.0.43
762
763- Fixed a bug where the theme selector was saving excessively
764- Hooks: Added EPIPE system error handling
765
766## 1.0.42
767
768- Added tilde (`~`) expansion support to `/add-dir` command
769
770## 1.0.41
771
772- Hooks: Split Stop hook triggering into Stop and SubagentStop
773- Hooks: Enabled optional timeout configuration for each command
774- Hooks: Added "hook_event_name" to hook input
775- Fixed a bug where MCP tools would display twice in tool list
776- New tool parameters JSON for Bash tool in `tool_decision` event
777
778## 1.0.40
779
780- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
781
782## 1.0.39
783
784- New Active Time metric in OpenTelemetry logging
785
786## 1.0.38
787
788- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
789
790## 1.0.37
791
792- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
793
794## 1.0.36
795
796- Web search now takes today's date into context
797- Fixed a bug where stdio MCP servers were not terminating properly on exit
798
799## 1.0.35
800
801- Added support for MCP OAuth Authorization Server discovery
802
803## 1.0.34
804
805- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
806
807## 1.0.33
808
809- Improved logging functionality with session ID support
810- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
811- Improvements to plan mode
812
813## 1.0.32
814
815- Updated loopback config for litellm
816- Added forceLoginMethod setting to bypass login selection screen
817
818## 1.0.31
819
820- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
821
822## 1.0.30
823
824- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
825- Improved file path autocomplete with filename matching
826- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
827- Enhanced jq regex support for complex filters with pipes and select
828
829## 1.0.29
830
831- Improved CJK character support in cursor navigation and rendering
832
833## 1.0.28
834
835- Slash commands: Fix selector display during history navigation
836- Resizes images before upload to prevent API size limit errors
837- Added XDG_CONFIG_HOME support to configuration directory
838- Performance optimizations for memory usage
839- New attributes (terminal.type, language) in OpenTelemetry logging
840
841## 1.0.27
842
843- Streamable HTTP MCP servers are now supported
844- Remote MCP servers (SSE and HTTP) now support OAuth
845- MCP resources can now be @-mentioned
846- /resume slash command to switch conversations within Claude Code
847
848## 1.0.25
849
850- Slash commands: moved "project" and "user" prefixes to descriptions
851- Slash commands: improved reliability for command discovery
852- Improved support for Ghostty
853- Improved web search reliability
854
855## 1.0.24
856
857- Improved /mcp output
858- Fixed a bug where settings arrays got overwritten instead of merged
859
860## 1.0.23
861
862- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
863- Released Python SDK: pip install claude-code-sdk to get started
864
865## 1.0.22
866
867- SDK: Renamed `total_cost` to `total_cost_usd`
868
869## 1.0.21
870
871- Improved editing of files with tab-based indentation
872- Fix for tool_use without matching tool_result errors
873- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
874
875## 1.0.18
876
877- Added --add-dir CLI argument for specifying additional working directories
878- Added streaming input support without require -p flag
879- Improved startup performance and session storage performance
880- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
881- Added detailed MCP server tools display (/mcp)
882- MCP authentication and permission improvements
883- Added auto-reconnection for MCP SSE connections on disconnect
884- Fixed issue where pasted content was lost when dialogs appeared
885
886## 1.0.17
887
888- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
889- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
890- MCP server list UI improvements
891- Update Claude Code process title to display "claude" instead of "node"
892
893## 1.0.11
894
895- Claude Code can now also be used with a Claude Pro subscription
896- Added /upgrade for smoother switching to Claude Max plans
897- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
898- Improved shell configuration error handling
899- Improved todo list handling during compaction
900
901## 1.0.10
902
903- Added markdown table support
904- Improved streaming performance
905
906## 1.0.8
907
908- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
909- Increased default otel interval from 1s -> 5s
910- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
911- Fixed a regression where search tools unnecessarily asked for permissions
912- Added support for triggering thinking non-English languages
913- Improved compacting UI
914
915## 1.0.7
916
917- Renamed /allowed-tools -> /permissions
918- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
919- Deprecated claude config commands in favor of editing settings.json
920- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
921- Improved error handling for /install-github-app
922- Bugfixes, UI polish, and tool reliability improvements
923
924## 1.0.6
925
926- Improved edit reliability for tab-indented files
927- Respect CLAUDE_CONFIG_DIR everywhere
928- Reduced unnecessary tool permission prompts
929- Added support for symlinks in @file typeahead
930- Bugfixes, UI polish, and tool reliability improvements
931
932## 1.0.4
933
934- Fixed a bug where MCP tool errors weren't being parsed correctly
935
936## 1.0.1
937
938- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
939- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
940- Updated documentation links and OAuth process descriptions
941
942## 1.0.0
943
944- Claude Code is now generally available
945- Introducing Sonnet 4 and Opus 4 models
946
947## 0.2.125
948
949- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
950- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
951
952## 0.2.117
953
954- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
955- Introduced settings.cleanupPeriodDays
956- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
957- Introduced --debug mode
958
959## 0.2.108
960
961- You can now send messages to Claude while it works to steer Claude in real-time
962- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
963- Fixed a bug where thinking was not working in -p mode
964- Fixed a regression in /cost reporting
965- Deprecated MCP wizard interface in favor of other MCP commands
966- Lots of other bugfixes and improvements
967
968## 0.2.107
969
970- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
971
972## 0.2.106
973
974- MCP SSE server configs can now specify custom headers
975- Fixed a bug where MCP permission prompt didn't always show correctly
976
977## 0.2.105
978
979- Claude can now search the web
980- Moved system & account status to /status
981- Added word movement keybindings for Vim
982- Improved latency for startup, todo tool, and file edits
983
984## 0.2.102
985
986- Improved thinking triggering reliability
987- Improved @mention reliability for images and folders
988- You can now paste multiple large chunks into one prompt
989
990## 0.2.100
991
992- Fixed a crash caused by a stack overflow error
993- Made db storage optional; missing db support disables --continue and --resume
994
995## 0.2.98
996
997- Fixed an issue where auto-compact was running twice
998
999## 0.2.96
1000
1001- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1002
1003## 0.2.93
1004
1005- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1006- Claude now has access to a Todo list that helps it stay on track and be more organized
1007
1008## 0.2.82
1009
1010- Added support for --disallowedTools
1011- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1012
1013## 0.2.75
1014
1015- Hit Enter to queue up additional messages while Claude is working
1016- Drag in or copy/paste image files directly into the prompt
1017- @-mention files to directly add them to context
1018- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1019- Improved performance for filename auto-complete
1020
1021## 0.2.74
1022
1023- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1024- Task tool can now perform writes and run bash commands
1025
1026## 0.2.72
1027
1028- Updated spinner to indicate tokens loaded and tool usage
1029
1030## 0.2.70
1031
1032- Network commands like curl are now available for Claude to use
1033- Claude can now run multiple web queries in parallel
1034- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1035
1036## 0.2.69
1037
1038- Fixed UI glitches with improved Select component behavior
1039- Enhanced terminal output display with better text truncation logic
1040
1041## 0.2.67
1042
1043- Shared project permission rules can be saved in .claude/settings.json
1044
1045## 0.2.66
1046
1047- Print mode (-p) now supports streaming output via --output-format=stream-json
1048- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1049
1050## 0.2.63
1051
1052- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1053
1054## 0.2.61
1055
1056- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1057- Enhanced image detection for more reliable clipboard paste functionality
1058- Fixed an issue where ESC key could crash the conversation history selector
1059
1060## 0.2.59
1061
1062- Copy+paste images directly into your prompt
1063- Improved progress indicators for bash and fetch tools
1064- Bugfixes for non-interactive mode (-p)
1065
1066## 0.2.54
1067
1068- Quickly add to Memory by starting your message with '#'
1069- Press ctrl+r to see full output for long tool results
1070- Added support for MCP SSE transport
1071
1072## 0.2.53
1073
1074- New web fetch tool lets Claude view URLs that you paste in
1075- Fixed a bug with JPEG detection
1076
1077## 0.2.50
1078
1079- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1080
1081## 0.2.49
1082
1083- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1084
1085## 0.2.47
1086
1087- Press Tab to auto-complete file and folder names
1088- Press Shift + Tab to toggle auto-accept for file edits
1089- Automatic conversation compaction for infinite conversation length (toggle with /config)
1090
1091## 0.2.44
1092
1093- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1094
1095## 0.2.41
1096
1097- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1098- MCP server startup no longer blocks the app from starting up
1099
1100## 0.2.37
1101
1102- New /release-notes command lets you view release notes at any time
1103- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1104
1105## 0.2.36
1106
1107- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1108- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1109
1110## 0.2.34
1111
1112- Vim bindings for text input - enable with /vim or /config
1113
1114## 0.2.32
1115
1116- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1117- Fix for some PersistentShell issues
1118
1119## 0.2.31
1120
1121- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1122- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1123
1124## 0.2.30
1125
1126- Added ANSI color theme for better terminal compatibility
1127- Fixed issue where slash command arguments weren't being sent properly
1128- (Mac-only) API keys are now stored in macOS Keychain
1129
1130## 0.2.26
1131
1132- New /approved-tools command for managing tool permissions
1133- Word-level diff display for improved code readability
1134- Fuzzy matching for slash commands
1135
1136## 0.2.21
1137
1138- Fuzzy matching for /commands